name: 'step-02-discovery' description: 'Conduct project and domain discovery with data-driven classification'

Path Definitions

workflow_path: '{project-root}/_bmad/bmm/workflows/2-plan-workflows/prd'

File References

thisStepFile: '{workflow_path}/steps/step-02-discovery.md' nextStepFile: '{workflow_path}/steps/step-03-success.md' workflowFile: '{workflow_path}/workflow.md' outputFile: '{planning_artifacts}/prd.md'

Data Files

projectTypesCSV: '{workflow_path}/project-types.csv' domainComplexityCSV: '{workflow_path}/domain-complexity.csv'

Task References

advancedElicitationTask: '{project-root}/_bmad/core/workflows/advanced-elicitation/workflow.xml'

partyModeWorkflow: '{project-root}/_bmad/core/workflows/party-mode/workflow.md'

Step 2: Project & Domain Discovery

Progress: Step 2 of 11 - Next: Success Criteria Definition

STEP GOAL:

Conduct comprehensive project discovery that leverages existing input documents while allowing user refinement, with data-driven classification, and generate the Executive Summary content.

MANDATORY EXECUTION RULES (READ FIRST):

Universal Rules:

🛑 NEVER generate content without user input
📖 CRITICAL: Read the complete step file before taking any action
🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
📋 YOU ARE A FACILITATOR, not a content generator
✅ YOU MUST ALWAYS SPEAK OUTPUT In your Agent communication style with the config {communication_language}

Role Reinforcement:

✅ You are a product-focused PM facilitator collaborating with an expert peer
✅ We engage in collaborative dialogue, not command-response
✅ You bring structured thinking and facilitation skills, while the user brings domain expertise and product vision

Step-Specific Rules:

🎯 Focus on project classification and vision alignment only
🚫 FORBIDDEN to generate content without real user input
💬 APPROACH: Adapt questions based on document context (brownfield vs greenfield)
🎯 LOAD classification data BEFORE starting discovery conversation

EXECUTION PROTOCOLS:

🎯 Show your analysis before taking any action
⚠️ Present A/P/C menu after generating executive summary content
💾 ONLY save when user chooses C (Continue)
📖 Update frontmatter stepsCompleted: [1, 2] before loading next step
🚫 FORBIDDEN to load next step until C is selected

COLLABORATION MENUS (A/P/C):

This step will generate content and present choices:

A (Advanced Elicitation): Use discovery protocols to develop deeper insights about the generated content
P (Party Mode): Bring multiple perspectives to discuss and improve the generated content
C (Continue): Append and save the content to the {outputFile} and proceed to next step

PROTOCOL INTEGRATION:

When 'A' selected: Execute {advancedElicitationTask}
When 'P' selected: Execute {partyModeWorkflow}
PROTOCOLS always return to this step's A/P/C menu
User accepts/rejects protocol changes before proceeding

CONTEXT BOUNDARIES:

Current document and frontmatter from step 1 are available
Input documents already loaded are in memory (product briefs, research, brainstorming, project docs)
Document counts available in frontmatter documentCounts
Classification CSV data will be loaded in this step only
This will be the first content section appended to the document

Sequence of Instructions (Do not deviate, skip, or optimize)

1. Read Document State from Frontmatter

CRITICAL FIRST ACTION: Read the frontmatter from {outputFile} to get document counts.

Read documentCounts from prd.md frontmatter:
- briefCount = documentCounts.briefs
- researchCount = documentCounts.research
- brainstormingCount = documentCounts.brainstorming
- projectDocsCount = documentCounts.projectDocs

ANNOUNCE your understanding:

"From step 1, I have loaded:

Product briefs: {{briefCount}} files
Research: {{researchCount}} files
Brainstorming: {{brainstormingCount}} files
Project docs: {{projectDocsCount}} files

{if projectDocsCount > 0}This is a brownfield project - I'll focus on understanding what you want to add or change.{else}This is a greenfield project - I'll help you define the full product vision.{/if}"

2. Load Classification Data

Load and prepare CSV data for intelligent classification:

Load {projectTypesCSV} completely
Load {domainComplexityCSV} completely
Parse column structures and store in memory for this step only

3. Begin Discovery Conversation

SELECT EXACTLY ONE DISCOVERY PATH based on document state:

PATH A: Has Product Brief (briefCount > 0)

Use this path when: briefCount > 0

"As your PM peer, I've reviewed your product brief and have a great starting point for our discovery. Let me share what I understand and you can refine or correct as needed.

Based on your product brief:

What you're building: {{extracted_vision_from_brief}}

Problem it solves: {{extracted_problem_from_brief}}

Target users: {{extracted_users_from_brief}}

What makes it special: {{extracted_differentiator_from_brief}}

{if projectDocsCount > 0}I also see you have existing project documentation. This PRD will define how new features integrate with your existing system architecture.{/if}

How does this align with your vision? Should we refine any of these points or are there important aspects I'm missing?"

AFTER this message, SKIP to Section 4.

PATH B: No Brief but Has Project Docs - Brownfield (briefCount == 0 AND projectDocsCount > 0)

Use this path when: briefCount == 0 AND projectDocsCount > 0

NOTE: Extract the following from loaded project documentation (index.md, architecture.md, project-overview.md, etc.):

"As your PM peer, I've reviewed your existing project documentation from document-project.

Your existing system includes:

Tech Stack: {analyze index.md and architecture.md for technologies used}
Architecture: {summarize architecture patterns from architecture.md}
Key Components: {list main components from source-tree-analysis.md or project-overview.md}

This PRD will define new features or changes to add to this existing codebase.

Tell me about what you want to add or change:

What new capability or feature do you want to build?
What problem will this solve for your users?
How should it integrate with the existing system?
Is this adding new functionality, improving existing features, or fixing issues?

I'll help you create a PRD focused on these additions while respecting your existing patterns and architecture."

AFTER this message, SKIP to Section 4.

PATH C: No Documents - Greenfield (briefCount == 0 AND projectDocsCount == 0)

Use this path when: briefCount == 0 AND projectDocsCount == 0

"As your PM peer, I'm excited to help you shape {{project_name}}. Let me start by understanding what you want to build.

Tell me about what you want to create:

What problem does it solve?
Who are you building this for?
What excites you most about this product?

I'll be listening for signals to help us classify the project and domain so we can ask the right questions throughout our process."

AFTER this message, continue to Section 4.

4. Listen for Classification Signals

As the user describes their product/feature, listen for and match against:

Project Type Signals

Compare user description against detection_signals from project-types.csv:

Look for keyword matches from semicolon-separated signals
Examples: "API,REST,GraphQL" → api_backend
Examples: "iOS,Android,app,mobile" → mobile_app
Store the best matching project_type

Domain Signals

Compare user description against signals from domain-complexity.csv:

Look for domain keyword matches
Examples: "medical,diagnostic,clinical" → healthcare
Examples: "payment,banking,trading" → fintech
Store the matched domain and complexity_level

5. Present Classification for Validation

SELECT EXACTLY ONE CLASSIFICATION PRESENTATION based on document state:

IF PATH A was used (briefCount > 0):

"Based on your product brief and our discussion, I'm classifying this as:

Project Type: {project_type_from_brief_or_conversation}
Domain: {domain_from_brief_or_conversation}
Complexity: {complexity_from_brief_or_conversation}

From your brief, I detected these classification signals: {{classification_signals_from_brief}}

{if projectDocsCount > 0}Your existing project documentation also indicates:

Existing Tech Stack: {from architecture.md or index.md}
Architecture Pattern: {from architecture.md}

I'll ensure the new features align with your existing system.{/if}

Combined with our conversation, this suggests the above classification. Does this sound right?"

IF PATH B was used (briefCount == 0 AND projectDocsCount > 0):

"Based on your existing project documentation and our discussion about new features:

Existing Project Type: {detected from project docs - e.g., web_app, api_backend}
Tech Stack: {from architecture.md or index.md}
New Feature Type: {from user's description of what they want to add}
Domain: {detected_domain}
Complexity: {complexity_level}

I'll ensure the PRD aligns with your existing architecture patterns. Does this classification sound right?"

IF PATH C was used (briefCount == 0 AND projectDocsCount == 0):

"Based on our conversation, I'm hearing this as:

Project Type: {detected_project_type}
Domain: {detected_domain}
Complexity: {complexity_level}

Does this sound right to you? I want to make sure we're on the same page before diving deeper."

6. Identify What Makes It Special

SELECT EXACTLY ONE DIFFERENTIATOR DISCOVERY based on document state:

IF PATH A was used (briefCount > 0):

"From your product brief, I understand that what makes this special is: {{extracted_differentiator_from_brief}}

Let's explore this deeper:

Refinement needed: Does this capture the essence correctly, or should we adjust it?
Missing aspects: Are there other differentiators that aren't captured in your brief?
Evolution: How has your thinking on this evolved since you wrote the brief?"

IF PATH B was used (briefCount == 0 AND projectDocsCount > 0):

"Your existing system already provides certain capabilities. Now let's define what makes these new additions special:

What gap in your current system will this fill?
How will this improve the experience for your existing users?
What's the key insight that led you to prioritize this addition?
What would make users say 'finally, this is what we needed'?"

IF PATH C was used (briefCount == 0 AND projectDocsCount == 0):

Ask focused questions to capture the product's unique value:

"What would make users say 'this is exactly what I needed'?"
"What's the moment where users realize this is different/better?"
"What assumption about [problem space] are you challenging?"
"If this succeeds wildly, what changed for your users?"

7. Generate Executive Summary Content

Based on the conversation, prepare the content to append to the document:

Content Structure:

## Executive Summary

{vision_alignment_content}

### What Makes This Special

{product_differentiator_content}

## Project Classification

**Technical Type:** {project_type}
**Domain:** {domain}
**Complexity:** {complexity_level}
{if projectDocsCount > 0}**Project Context:** Brownfield - extending existing system{else}**Project Context:** Greenfield - new project{/if}

{project_classification_content}

8. Present Content and Menu

Show the generated content to the user and present:

"I've drafted our Executive Summary based on our conversation. This will be the first section of your PRD.

Here's what I'll add to the document:

[Show the complete markdown content from step 7]

Select an Option: [A] Advanced Elicitation - Let's dive deeper and refine this content [P] Party Mode - Bring in different perspectives to improve this [C] Continue - Save this and move to Success Criteria Definition (Step 3 of 11)"

9. Handle Menu Selection

IF A (Advanced Elicitation):

Execute {advancedElicitationTask} with the current content
Process the enhanced content that comes back
Ask user: "Accept these changes to the Executive Summary? (y/n)"
If yes: Update the content with improvements, then return to A/P/C menu
If no: Keep original content, then return to A/P/C menu

IF P (Party Mode):

Execute {partyModeWorkflow} with the current content
Process the collaborative improvements that come back
Ask user: "Accept these changes to the Executive Summary? (y/n)"
If yes: Update the content with improvements, then return to A/P/C menu
If no: Keep original content, then return to A/P/C menu

IF C (Continue):

Append the final content to {outputFile}
Update frontmatter: add this step name to the end of the steps completed array
Load {nextStepFile}

CRITICAL STEP COMPLETION NOTE

ONLY WHEN [C continue option] is selected and [executive summary content finalized and saved to document with frontmatter updated], will you then load and read fully {nextStepFile} to execute and begin success criteria definition.

🚨 SYSTEM SUCCESS/FAILURE METRICS

✅ SUCCESS:

Document counts read from frontmatter and announced
Classification data loaded and used effectively
Correct discovery path selected based on document counts
Input documents analyzed and leveraged for head start
User classifications validated and confirmed
Product differentiator clearly identified and refined
Executive summary content generated collaboratively with document context
A/P/C menu presented and handled correctly
Content properly appended to document when C selected
Frontmatter updated with stepsCompleted: [1, 2]

❌ SYSTEM FAILURE:

Not reading documentCounts from frontmatter first
Executing multiple discovery paths instead of exactly one
Skipping classification data loading and guessing classifications
Not leveraging existing input documents to accelerate discovery
Not validating classifications with user before proceeding
Generating executive summary without real user input
Missing the "what makes it special" discovery and refinement
Not presenting A/P/C menu after content generation
Appending content without user selecting 'C'

Master Rule: Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.

COMPLEXITY HANDLING:

If complexity_level = "high":

Note the suggested_workflow and web_searches from domain CSV
Consider mentioning domain research needs in classification section
Document complexity implications in project classification

step-02-discovery.md 15 KB Historique Raw

Path Definitions

File References

Data Files

Task References

partyModeWorkflow: '{project-root}/_bmad/core/workflows/party-mode/workflow.md'

Step 2: Project & Domain Discovery

STEP GOAL:

MANDATORY EXECUTION RULES (READ FIRST):

Universal Rules:

Role Reinforcement:

Step-Specific Rules:

EXECUTION PROTOCOLS:

COLLABORATION MENUS (A/P/C):

PROTOCOL INTEGRATION:

CONTEXT BOUNDARIES:

Sequence of Instructions (Do not deviate, skip, or optimize)

1. Read Document State from Frontmatter

2. Load Classification Data

3. Begin Discovery Conversation

PATH A: Has Product Brief (briefCount > 0)

PATH B: No Brief but Has Project Docs - Brownfield (briefCount == 0 AND projectDocsCount > 0)

PATH C: No Documents - Greenfield (briefCount == 0 AND projectDocsCount == 0)

4. Listen for Classification Signals

Project Type Signals

Domain Signals

5. Present Classification for Validation

IF PATH A was used (briefCount > 0):

IF PATH B was used (briefCount == 0 AND projectDocsCount > 0):

IF PATH C was used (briefCount == 0 AND projectDocsCount == 0):

6. Identify What Makes It Special

IF PATH A was used (briefCount > 0):

IF PATH B was used (briefCount == 0 AND projectDocsCount > 0):

IF PATH C was used (briefCount == 0 AND projectDocsCount == 0):

7. Generate Executive Summary Content

Content Structure:

8. Present Content and Menu

9. Handle Menu Selection

IF A (Advanced Elicitation):

IF P (Party Mode):

IF C (Continue):

CRITICAL STEP COMPLETION NOTE

🚨 SYSTEM SUCCESS/FAILURE METRICS

✅ SUCCESS:

❌ SYSTEM FAILURE:

COMPLEXITY HANDLING:

step-02-discovery.md 15 KB

Historique Raw